Improving the Accuracy of Daily Satellite-Derived Ground-Level Fine Aerosol Concentration Estimates for North America

摘要

We improve the accuracy of daily ground-level fine partalate matter concentrations (PM_(2.5)) derived from satellite observations (MODIS and MISR) of aerosol optical depth (AOD) and chemical transport model (GEOS-Chem) calculations of the relationship between AOD and PM_(2.5). This improvement is achieved by (1) applying dimatological ground-based regional bias-correction factors based upon comparison with in situ PM_(2.5), and (2) applying spatial smoothing to reduce random uncertainty and extend coverage. Initial daily 1-σ mean uncertainties are reduced across the United States and southern Canada from ± (1 μg/m~3 + 67%) to ± (1 μg/m~3 + 54%) by applying the dimatological ground-based regional scaling factors. Spatial interpolation increases the coverage of satellite-derived PM_(2.5) estimates without increased uncertainty when in close proximity to direct AOD retrievals. Spatial smoothing further reduces the daily 1-σ uncertainty to ±(1 μg/m~3 + 42%) by limiting the random component of uncertainty. We additionally find similar performance for dimatological relationships of AOD to PM_(2.5) as compared to day-specinc relationships.